Uploaded image for project: 'MariaDB Server'
  1. MariaDB Server
  2. MDEV-14875

galera_new_cluster crashes mysqld when existing server contains databases

Details

    • Bug
    • Status: Closed (View Workflow)
    • Critical
    • Resolution: Fixed
    • 10.2.12
    • 10.2.13
    • Debian 9 fresh install
      mariadb from debian repos
      galera from debian repos
    • 10.2.13

    Description

      Try to setup galera cluster from standalone mariadb server that has a few databases.

      From existing mariadb server use /var/lib/mysql as is
      on stock /etc/mysql/my.cnf
      add

      [galera]

      1. Mandatory settings
        wsrep_on=ON
        wsrep_provider=/usr/lib/galera/libgalera_smm.so
        wsrep_cluster_name="tam_cluster"
        wsrep_cluster_address="gcomm://ip1, ip2, ip5"
        wsrep_node_address="ip5"
        wsrep_node_name="node_5"
        wsrep_sst_method=rsync
        binlog_format=row
        default_storage_engine=InnoDB
        innodb_autoinc_lock_mode=2

      run: galera_new_cluster

      result: mysqld crashes

      Attachments

        1. jgnoss_first_pages_001.bin
          128 kB
        2. mdb_dbg.tgz
          302 kB
        3. my.cnf
          6 kB
        4. mysqld_debug.txt
          17 kB
        5. mysqld.debug_outputs.txt
          14 kB
        6. mysqld.trace
          9.42 MB
        7. mysqld.trace
          1.76 MB
        8. mysqld.trace
          1.82 MB
        9. mysqld.trace
          88 kB

        Activity

          jgnoss Juergen Gnoss added a comment -

          I just added the trace of the last

          galera_new_cluster

          jgnoss Juergen Gnoss added a comment - I just added the trace of the last galera_new_cluster

          Can I have full error log? Something must go wrong at wsrep initialization.

          jplindst Jan Lindström (Inactive) added a comment - Can I have full error log? Something must go wrong at wsrep initialization.
          jgnoss Juergen Gnoss added a comment -

          here the files from an actual try to bootstrap the cluster

          on my debian 9 install if I use galera_new_cluster the log goes to syslog
          so the attached file has the relevant output from syslog

          @jan: would it be of help to chat on freenode #maria ?

          jgnoss Juergen Gnoss added a comment - here the files from an actual try to bootstrap the cluster on my debian 9 install if I use galera_new_cluster the log goes to syslog so the attached file has the relevant output from syslog @jan: would it be of help to chat on freenode #maria ?

          Yes, it would help, what files you have on datadir ? You should not have galera state files when you create a new cluster, why it can't create normal error log?

          jplindst Jan Lindström (Inactive) added a comment - Yes, it would help, what files you have on datadir ? You should not have galera state files when you create a new cluster, why it can't create normal error log?

          It seems that MariaDB 10.2.13 has fixed this issue:

          jan@jan-laptop-asus:/$ sudo /usr/local/bin/galera_new_cluster 
          jan@jan-laptop-asus:/$ sudo systemctl status mariadb
          ● mariadb.service - MariaDB 10.2.13 database server
             Loaded: loaded (/lib/systemd/system/mariadb.service; enabled; vendor preset: enabled
            Drop-In: /etc/systemd/system/mariadb.service.d
                     └─migrated-from-my.cnf-settings.conf
             Active: active (running) since Tue 2018-02-06 16:41:22 EET; 1 day 18h ago
               Docs: man:mysqld(8)
                     https://mariadb.com/kb/en/library/systemd/
           Main PID: 2316 (mysqld)
             Status: "Taking your SQL requests now..."
              Tasks: 35 (limit: 4915)
             CGroup: /system.slice/mariadb.service
                     └─2316 /usr/sbin/mysqld --wsrep_start_position=00000000-0000-0000-0000-00000
           
          Warning: Journal has been rotated since unit was started. Log output is incomplete or u
          

          jplindst Jan Lindström (Inactive) added a comment - It seems that MariaDB 10.2.13 has fixed this issue: jan@jan-laptop-asus:/$ sudo /usr/local/bin/galera_new_cluster jan@jan-laptop-asus:/$ sudo systemctl status mariadb ● mariadb.service - MariaDB 10.2.13 database server Loaded: loaded (/lib/systemd/system/mariadb.service; enabled; vendor preset: enabled Drop-In: /etc/systemd/system/mariadb.service.d └─migrated-from-my.cnf-settings.conf Active: active (running) since Tue 2018-02-06 16:41:22 EET; 1 day 18h ago Docs: man:mysqld(8) https://mariadb.com/kb/en/library/systemd/ Main PID: 2316 (mysqld) Status: "Taking your SQL requests now..." Tasks: 35 (limit: 4915) CGroup: /system.slice/mariadb.service └─2316 /usr/sbin/mysqld --wsrep_start_position=00000000-0000-0000-0000-00000   Warning: Journal has been rotated since unit was started. Log output is incomplete or u

          People

            jplindst Jan Lindström (Inactive)
            jgnoss Juergen Gnoss
            Votes:
            0 Vote for this issue
            Watchers:
            5 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved:

              Git Integration

                Error rendering 'com.xiplink.jira.git.jira_git_plugin:git-issue-webpanel'. Please contact your Jira administrators.